A perceptual discrimination task abnormally …nzbri.org/resources/publications/60/van_Stockum_European...A perceptual discrimination task abnormally facilitates reﬂexive saccades

A perceptual discrimination task abnormally facilitatesreflexive saccades in Parkinson’s disease

Saskia van Stockum,1,2 Michael R. MacAskill,1,2 Daniel Myall1 and Tim J. Anderson1,2,3

1Van der Veer Institute for Parkinson’s & Brain Research, 66 Stewart Street, Christchurch, New Zealand2Department of Medicine, University of Otago, Christchurch, New Zealand3Department of Neurology, Christchurch Hospital, Christchurch, New Zealand

Keywords: human, Parkinson’s disease, saccades, visual attention

Abstract

Numerous studies have shown that Parkinson’s disease (PD) affects the ability to generate voluntary saccades and the ability tosuppress reflexive saccades. The effects of PD on the generation of reflexive saccades, however, are not clear. Some studies reportimpairments, but there are also reports of abnormal facilitation or hyper-reflexivity of the saccade system in PD. Meanwhile, it hasbeen reported that the concurrent performance of a perceptual discrimination task facilitates saccade initiation and reduces saccadelatencies in healthy subjects [A. Montagnini & L. Chelazzi (2005) Vis. Res., 45, 3391–3401; L. Trottier & J. Pratt (2005) Vis. Res., 45,1349–1354]. To investigate the circumstances under which the saccade system may appear hyper-reflexive in PD, we comparedreflexive saccades with and without a concurrent perceptual discrimination task in 20 PD patients and 20 controls. Without thediscrimination task, the PD group produced reflexive saccades at normal latencies. The discrimination task reduced saccadelatencies more in the PD group than in the control group, resulting in abnormally short mean reflexive saccade latencies in the PDgroup. The discrimination task increased saccade gain in both groups, but saccades in the PD group remained hypometric ascompared with saccades in the control group. We conclude that the attentional demands of this paradigm revealed a hypersensitivityto visual inputs in the PD group.

Introduction

In Parkinson’s disease (PD), the function of the saccadic system isaffected by dopamine depletion in the basal ganglia. Impairments inPD have been consistently detected in tasks that require the generationof voluntary saccades, such as delayed, memory-guided or anti-saccade tasks. These tasks require, in addition to the execution ofvoluntary saccades, the inhibition of reflexive saccades. PD patients’impairments in these tasks include hypometria (the eyes initially landshort of the target, and some catch-up steps are required to foveatethe intended location) (Lueck et al., 1992; Shaunak et al., 1999;Armstrong et al., 2002; Le Heron et al., 2005), prolonged latencies,and failure to suppress unwanted reflexive saccades towards a visualstimulus (Briand et al., 1999; Armstrong et al., 2002; Chan et al.,2005; Amador et al., 2006; Gurvich et al., 2007). In contrast, reportson the performance of reflexive saccade tasks in PD are inconsistent[see Chambers & Prescott (2010) for a review]. Reflexive saccades areoften thought to be normal in PD, but there are also reports ofprolonged latency (Chen et al., 1999) and of abnormal facilitation ofthe reflexive saccadic system (or hyper-reflexivity; Briand et al., 2001;

Armstrong et al., 2002; Kingstone et al., 2002; Chan et al., 2005; vanStockum et al., 2008).The tasks most often used to investigate voluntary saccades involve,

in addition to the programming of a voluntary saccade, the suppres-sion of a reflexive saccade. That is, subjects must shift attention to avisual stimulus, without making a saccade to that stimulus. In theseparadigms, top-down saccade selection competes with bottom-upvisual input. Performance of these tasks will be impaired if top-downcontrol of the saccade system is impaired, but also if bottom-upprocesses are abnormally active.Current models of PD propose that progressive degeneration of

dopaminergic inputs into the striatum causes overactivity of inhibitoryoutputs from the basal ganglia via the substantia nigra pars reticulata(SNr) to the superior colliculus and via the thalamus to the cortex(Mink, 1996; Hikosaka et al., 2000). The basal ganglia are cruciallyinvolved in the selection of voluntary saccades and in the suppressionof unwanted saccades (Hikosaka et al., 2000). Excessive inhibitoryoutput from the basal ganglia in the saccade system in PD is thought toaffect the ability to generate voluntary saccades and the ability tosuppress unwanted reflexive saccades (Chan et al., 2005; Gurvichet al., 2007; Hood et al., 2007; Cameron et al., 2010). In general,impairments of the saccade system in PD have been interpreted asevidence of a failure of top-down control, owing to disruption offronto-striatal circuitry. However, it is not clear why hyper-reflexivityis observed only in some studies, and whether hyper-reflexivity (when

Correspondence: S. van Stockum, 1Van der Veer Institute for Parkinson’s & BrainResearch, as above.E-mail: [email protected]

Received 30 November 2010, revised 15 February 2011, accepted 16 March 2011

European Journal of Neuroscience, Vol. 33, pp. 2091–2100, 2011 doi:10.1111/j.1460-9568.2011.07697.x

ª 2011 The Authors. European Journal of Neuroscience ª 2011 Federation of European Neuroscience Societies and Blackwell Publishing Ltd

European Journal of Neuroscience

it does occur) should be attributed to a deficit of top-down control inPD, or to overactivity of bottom-up processes.To further investigate the circumstances under which the reflexive

saccade system may be abnormally facilitated in PD, we adapted anexperimental paradigm in which saccades are performed together witha perceptual discrimination task (Deubel, 2008). The instruction tomake a perceptual discrimination at the saccade target has been foundto facilitate saccade initiation and reduce saccade latencies in healthysubjects (Montagnini & Chelazzi, 2005; Trottier & Pratt, 2005). Thecombination of a discrimination task with a saccade task allows theinvestigation of top-down and bottom-up influences in the saccadesystem, without competition between top-down saccade control andbottom-up visual inputs. Instead of evoking competition, the demandsof the discrimination task enhance the performance of the saccade taskby promoting a shift of attention to the saccade target. The use of thistask also enabled us to investigate the relationship between saccadicand visuospatial attention deficits in PD. Comparison with the studyusing the original version of this paradigm (Deubel, 2008) will not beappropriate, because that study did not use reflexive saccades, and thesaccade task was not performed without the discrimination task.

Materials and methods

Participants

Two groups were recruited: 20 PD patients (eight females) and 20control subjects (eight females). The groups were matched for meanage and years of education. The mean age in the PD group was65.0 years, ranging from 50 to 77 years [Hoehn and Yahr (H&Y)score 1–3]. In the control group, the mean age was 65.5 years, rangingfrom 56 to 76 years. Only subjects who scored 25 or more on theMontreal Cognitive Assessment (Nasreddine et al., 2005) wereincluded. The Movement Disorder Society-sponsored revision of theUnified Parkinson’s Disease Rating Scale (MDS-UPDRS) was used toassess motor impairment in the PD group (Goetz et al., 2008). Thesubjects in the PD group were tested ‘on’ medication. This projectreceived ethical approval from the Upper South A Regional Ethics

Committee. All participants gave informed consent. See Table 1 fordetails of the subjects in the PD group.

Apparatus and stimuli

Eye movements were recorded monocularly with a video-based iViewX Hi-Speed system (SMI, Berlin, Germany) at a sampling rate of1250 Hz. This system uses a combination of corneal reflection andpupil tracking, with a typical spatial accuracy of 0.25–0.5� and atracking resolution of < 0.01�. Stimuli were displayed on a 21-inchCRT screen with a 100-Hz refresh rate on a display area of400 · 300 mm, at a resolution of 800 · 600 pixels. The computerscreen was positioned 600 mm in front of the subject, who sat with thehead supported by the chin and forehead rest of the iView trackingcolumn. As PD patients may have lower contrast sensitivity and asmaller ‘useful field of view’ than control subjects (Uc et al., 2005),high-contrast stimuli and small target amplitude were used tominimize any potential differences in perceptual ability between thegroups. Stimuli were presented on a dark grey background (R50 G50B50). The fixation point was a red (R255 G0 B0) square(0.67 · 0.67�); targets (0.62� wide · 1� high) were white(R255 G255 B255). Targets appeared at the four corners of animaginary square, each 5.4� from the central fixation point. Each blockof trials started with a check of the calibration quality and, if required,a two-dimensional 13-point recalibration procedure covering thedisplay area was conducted. At the beginning and end of eachrecording, a sequence of reflexive saccades was recorded to providedata for post hoc assessment and adjustment of the calibration ifrequired. Stimuli were presented with PsychoPy, an open-sourceexperimental-control software package (Peirce, 2007, 2008).

Saccade and perceptual discrimination tasks

The paradigm, adapted from Deubel (2008), is one in which saccadescan be performed, with or without a concurrent perceptual discrim-ination task. In our study, the onset of a figure 8 was used as the cue

Table 1. Details of the subjects in the PD group

Age(years) Sex

Years ofeducation

Yearswith PD MoCA

MDS-UPDRSPart III H&Y Medications

50 F 14 4 26 29 3 Amantadine54 M 16 3 30 25 2 Ropinirole, amantadine58 M 14 13 30 54 2.5 Selegiline, Sinemet, Dopergin, amantadine59 M 12 2 30 22 1 Ropinirole, domperidone60 M 12 2 30 18 1 Selegeline, Sinemet62 M 16 2 26 30 2 Amantadine63 F 8 3 29 22 2 Selegiline, Madopar, Sinemet63 M 15 2 29 22 1 Moclobemide, Sinemet65 F 10 5 30 33 1.5 Ropinirole, amantadine65 M 10 1 28 31 2 Ropinirole, amitryptiline, amantadine65 M 12 4 25 22 2 Ropinirole, Sinemet, amantadine67 M 10 2 27 40 2 Sinemet, benztropine, amantadine68 F 14 3 27 15 1 None68 F 9 3 27 36 1 Amantadine70 F 11 14 27 17 2 Madopar, pergolide, amantadine70 F 15 9 26 55 2.5 Sinemet, ropinirole, amantadine70 M 18 7 30 44 2 Madopar, entacapone, amantadine72 F 15 9 25 25 1 Sinemet, fluoxetine74 M 10 2 30 38 2 Sinemet77 M 18 4 27 70 2.5 Madopar

F, female; M, male; MoCA, Montreal Cognitive Assessment. The MoCA is a general cognitive screening test. The MDS-UPDRS Part III is a scale assessing themotor signs of PD. The H&Y score reflects disease severity in PD.

2092 S. van Stockum et al.

ª 2011 The Authors. European Journal of Neuroscience ª 2011 Federation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience, 33, 2091–2100

for a reflexive saccade. The perceptual discrimination task requiredsubjects to report the identity of a symbol, which appeared briefly atthe target location shortly after target onset. Subjects first performedthe saccade task without the perceptual discrimination task. In thiscondition, the instruction was: ‘Move your eyes as quickly as possibleto the target’. Subsequently, subjects performed the saccade task withthe perceptual discrimination task. In this condition, the instructionwas: ‘Move your eyes as quickly as possible to the target, and reportafter each trial whether an E or 3 was displayed’. After making asaccade, subjects responded with a manual button press to indicatewhether they had seen an E (right button) or a 3 (left button).Instructions stressed the importance of making a saccade on each trial.Also, it was made clear that it was important to guess an answer evenif the subject was not sure which symbol had been displayed, and topush a button at random after trials where the subject thought nosymbol had appeared.

Stimulus sequence

Each trial started with a variable fixation interval. Targets consisted offigure 8s, which changed briefly (for 100 ms) into either E or 3 at 25,50, 100 or 150 ms after target onset [the stimulus onset asynchrony(SOA)]. In gap trials, the fixation point disappeared 200 ms beforetarget onset, and in overlap trials, the fixation point remained until theend of the display of the discrimination symbol. In distractor trials, a

Fig. 2. Mean latencies are shown for saccades, made by each group, with andwithout a concurrent discrimination task. Gap trials are shown on the left, andoverlap trials on the right. The error bars represent 95% confidence intervals.The concurrent performance of the perceptual task reduced saccade latenciesmore in the PD group than in the control group.

Fig. 1. Schematic illustration of the stimulus display, showing the sequence of events during each trial. Two gap conditions (with a gap and with an overlap) andtwo trial types (with and without a distractor) were used. In gap trials, the fixation point disappeared 200 ms before target onset, and in overlap trials, the fixationpoint remained visible until the discrimination symbol at the target location disappeared. In distractor trials, a Figure 2 or 5 appeared at a location diagonally oppositethe target location, simultaneously with the appearance of the discrimination symbol at the target location (see the shaded column in the centre). The sequence ofevents was identical for the saccade task without and the saccade task with the perceptual discrimination task; only the instructions to the subjects differed.

Abnormal facilitation of saccades in PD 2093


Figures 2 or 5 appeared at the location diagonally opposite the targetlocation for 100 ms, simultaneously with the discrimination symbolonset (Fig. 1). Target locations, gap conditions, trials with and withoutdistractor and the two discrimination symbols were balanced acrosstrial blocks. During each block of trials, three of the four targetlocations (top right, bottom left, top left, or bottom right) were used ateach SOA (25, 50, 100 or 150 ms), in each trial type (with or withoutdistractor) and in each gap condition (gap or overlap). Four differenttrial blocks were used to balance the target locations used at eachSOA. On half of these trials, the discrimination symbol was 3, and onthe other half it was E. Interspersed in each block of trials were eighttrials with target onsets, but without symbol changes. Each blockconsisted of 56 trials. For the analysis, the data from all trials werepooled across the four SOAs. The effects of SOA are not relevant tothe present report and are not reported here.

Procedure

All subjects attended two testing sessions, 1 week apart. Vision wastested at the start of each session, and corrected to normal withspectacles if necessary. Subjects controlled the timing of the trials bypushing a button to start each new trial only when they were ready,and looking at the central fixation point. Practice trials were presented,allowing subjects to become familiar with the push buttons and thetask requirements before the start of the actual test. In the first session,two blocks of trials were performed ‘without discrimination’, andsubjects were told to concentrate on moving their eyes as quickly aspossible to the targets. It was explained that they might see somethingflicker in the display on some trials, but that this was irrelevant to thetask. Next, two blocks of trials were performed ‘with discrimination’.Now, subjects were asked to pay attention to the symbol changes at thetarget location, because after each trial they would be asked to identifywhich symbol (3 or e) was shown. After each of these saccade trials, aprompt appeared on the monitor, asking the subject to report whichsymbol had appeared. In the second session, two more blocks of trials‘with discrimination’ were performed. Each block of trials waspresented in a different randomized order.

Measurement and analysis

Latency and gain of the primary saccade (the first saccade following theappearance of the target) were measured off-line. Latency wascalculated from the onset of the target. Gain was defined as theabsolute amplitude of the saccade divided by the target vector (whichwas always 5.4�). The eye position trace was searched to find the firstinstance where the eye velocity reached 80� ⁄ s after cue onset. Thebeginning and end of a primary saccade were determined by searchingbackwards and forwards in time from this point for the nearest velocityminimum. Trials without evidence of a saccade (because of blinks or no

response), with saccades at latencies shorter than 80 ms or longer than600 ms, or with a gain of less than 0.3 or more than 1.3, were excluded.Trials with saccades initiated at latencies longer than 80 ms, but withdirectional errors (the primary saccade was not directed at the cuedtarget location), were analysed separately. The proportion of excludedtrials did not differ between the control group and the PD group,leaving 83 and 81% of trials for analysis from each group respectively.The proportion of correct discriminations was calculated from the totalnumber of remaining trials with symbol onsets in the blocks of trialswith the perceptual discrimination task. Mean latencies and gain ofsaccades made in the trials with the discrimination task did not differbetween the first and the second session in either group, so the resultsfrom the two sessions were pooled. Effects of the discrimination taskon saccade latency and gain were analysed with linear mixed-effectsmodels. The function lme from the r package nlme was used to fit themodels (Pinheiro, Bates, DebRoy, Sarkar, & Team, 2009). For piece-wise linear regression, bentcableAR was used (Chiu, 2010).

Results

Latencies

Latencies were analysed with a linear mixed-effects model, withGroup as a between-subjects factor (control or PD), and Task (with orwithout discrimination task), Gap (gap or overlap) and Trial type (withor without distractor) as within-subjects factors. Effects are shown asmean values followed by 95% confidence intervals.The Gap manipulation had a significant effect on saccade latencies:

the overlap condition lengthened mean saccade latency by 97 ms (84,111) as compared with trials with a gap (t39 = 14.59, P < 0.001).There was also a main effect of Trial type: distractors lengthened meansaccade latency by 29 ms (34, 24) (t159 = 12.25, P < 0.001). Overall,there was a significant interaction between Gap and Task(F1, 77 = 10.35, P < 0.01): in trials with a gap, the discriminationtask did not affect saccade latencies, but in trials with an overlap, thediscrimination task shortened latencies by 29 ms ()47, )11)(t77 = )3.22, P < 0.01). Overall, there was also a significant interac-tion between Group and Task (F1, 77 = 4.56, P < 0.05): the discrim-ination task reduced latencies more in the PD group than in the controlgroup. No other interactions were significant. Figure 2 shows theeffects on saccade latency pooled across trials with and withoutdistractors. The mean latencies of reflexive saccades, in gap andoverlap trials, with and without the concurrent discrimination task areshown for each group in Table 2.

Production of express saccades

Proportions of express saccades were compared with a mixed-effectsbinomial model with Group (control or PD), Task (saccades with or

Table 2. Mean saccade latency (ms; 95% confidence interval) in gap and overlap trials, pooled across trials with and trials without a distractor, in which saccadeswere made with and without the perceptual discrimination task

Gap trials Overlap trials

Withoutdiscrimination

Withdiscrimination


Withdiscrimination

Control 223 (201, 246) 231 (209, 253) 323 (301, 346) 307 (285, 329)PD 212 (189, 235) 203 (181, 226) 310 (287, 332) 267 (245, 290)*

*Difference between the groups, P < 0.05.



without discrimination task), Gap (gap or overlap) and Trial type (withor without distractor) as factors. There was an effect of Group: the PDgroup made more express saccades than the control group (14% vs.5%, z = 2.16, P = 0.03). There was an effect of Gap: in overlap trials,the proportion of express saccades was smaller than in gap trials(z = )10.29, P < 0.001). There was also an interaction betweenGroup and Gap. The overlap decreased the production of expresssaccades more in the control group (from 10 to 1%) than in the PDgroup (from 21 to 7%; z = 2.72, P < 0.01). The discrimination task orthe distractors did not affect the production of express saccades ineither group. Table 3 shows the proportions of express saccades foreach group, in trials with and without a distractor, in gap and overlaptrials, with and without the discrimination task.

Primary saccade gain

Saccade gain was compared with a linear mixed-effects model, withGroup as a between-subjects factor (control or PD), and Task (with orwithout discrimination task), Gap (gap or overlap) and Trial type (withor without distractor) as within-subjects factors. Effects are shown asmean values followed by 95% confidence intervals. There were maineffects of Group, Gap, Trial type and Task on the gain of saccades.The mean gain of saccades in the PD group was 0.10 ()0.15, )0.05)smaller than in the control group (t38 = )4.33, P < 0.001). In overlaptrials, saccades were 0.03 (0.01, 0.05) larger than in gap trials(t39 = 3.17, P < 0.01). Distractors decreased mean saccade gain by0.02 ()0.01, )0.03) (t159 = )4.29, P < 0.001). The discriminationtask increased saccade gain by 0.02 (0.001, 0.04) (t77 = 2.06,P < 0.05). There were no significant interactions. Figure 3 showsthe effects on saccade gain pooled across trials with and withoutdistractors. Mean gain values of reflexive saccades, in gap and overlaptrials, with and without the concurrent discrimination task, are shownfor each group in Table 4.

Directional errors

All trials with saccades initiated at latencies between 80 and 600 mswere included in the following analysis. For each subject, theproportion of trials with directional errors (saccades that were notdirected at the target location) was calculated. Proportions ofdirectional errors were compared with a mixed-effects binomialmodel, with Group (control or PD), Gap (gap or overlap), Task(saccades with or without discrimination task) and Trial type (with orwithout distractor) as factors. There were main effects of Gap, Task,and Trial type. More errors were made in trials with a gap than in trialswith an overlap (z = 4.48, P < 0.001). More errors were made in trialswith the discrimination task than in trials without the discrimination

task (z = )2.08, P < 0.05). More errors were made in trials with adistractor than in trials without a distractor (z = )14.18, P < 0.001).There were no significant interactions. Only 3% of directional errors intrials with distractors were not directed at the distractor location.Table 5 shows the proportions of directional errors for each group ineach condition.

Perceptual discrimination task

Proportions of correct judgements were compared with a mixed-effects binomial model, with Group (control or PD), Gap (gap oroverlap) and Trial type (with or without distractor) as factors. Therewas a main effect of Group. The control group made more correctjudgements than the PD group (79% vs. 68%, z = )3.00, P < 0.01).There was also a main effect of Trial type: fewer correct judgementswere made in trials with a distractor than in trials without a distractor(z = )2.31, P = 0.02). There were no significant interactions. Table 6shows the proportion of correct discrimination judgements in eachgroup.

Table 3. Proportion of express saccades (%) in the tasks with and without the concurrent discrimination task in trials with and in trials without a distractor in eachgroup

Without distractor With distractor


Withdiscrimination


Withdiscrimination

Gap Overlap Gap Overlap Gap Overlap Gap Overlap

Control 9 1 9 2 12 1 9 1PD 20 7* 21* 9* 24 5 21 8*


Fig. 3. Mean gain values are shown for primary saccades, made by eachgroup, with and without a concurrent discrimination task. Gap trials are shownon the left, and overlap trials on the right. The error bars represent 95%confidence intervals. The concurrent performance of the perceptual discrim-ination task resulted in increased saccade gain in the PD group, but saccadesremained hypometric as compared with the control group.



The proportion of trials in which the saccade was initiated before,during or after the display of the discrimination symbol was calculatedfor each group. These proportions did not differ between the groups.Table 7 shows these proportions, together with the proportion ofcorrectly identified discrimination symbols in each category.

Association between saccade production and discriminationperformance

In the control group, the performance of the discrimination task wasassociated with mean saccade latency: subjects who had longer meansaccade latencies made more correct discrimination judgements(r = 0.52, P < 0.01). In the PD group, the performance of thediscrimination task was associated with mean saccade gain: subjectswho had larger mean gain values made more correct discriminationjudgements (r = 0.57, P < 0.01); see Fig. 4.

Association between saccade latency and gain

In both groups, longer latencies resulted in saccades with higher gainvalues, but only when saccades were initiated at relatively short

latencies. A piece-wise linear model showed that, on average, gainincreased linearly with latency, until, by a latency of 176 ms, gainremained constant (Fig. 5). Two separate mixed-effects models wereused to test effects on the gain of short-latency (below 176 ms)saccades and of long-latency (over 176 ms) saccades. For saccadeswith latencies below 176 ms, there were main effects of Latency,Group, and Task. Overall, for saccades initiated at latencies below176 ms, gain increased from the intercept at 0.57 by 0.03 (0.02, 0.04)per 10 ms of increase in latency (t1641 = 7.49, P < 0.001). The PDgroup made saccades at latencies below 176 ms that were 0.06 (0.13,0.005) smaller than in the control group (t38 = 2.33, P = 0.03).Overall, the concurrent performance of the perceptual discriminationtask increased saccade gain by 0.04 (0.01, 0.07) (t64 = 2.51,P = 0.01). No interactions were significant.When the gain of saccades with latencies longer than 176 ms was

analysed, the main effect of Group remained: the mean gain value ofthese longer-latency saccades in the PD group was 0.06 (0.10, 0.01)smaller than in the control group (t38 = )3.77, P < 0.001). In contrastto the gain of short-latency saccades, the gain of saccades withlatencies longer than 176 ms basically remained constant, andincreased by only 0.0007 (0.0005, 0.0013) per 10 ms of latency.Overall, gain was only marginally increased by the perceptual

Table 4. Mean saccade gain (95% confidence interval) in gap and overlap trials, pooled across trials with and trials without a distractor, in which saccades weremade with and without the perceptual discrimination task



Withdiscrimination


Withdiscrimination

Control 0.86 (0.83, 0.90) 0.89 (0.86, 0.92) 0.90 (0.87, 0.94) 0.90 (0.86, 0.93)PD 0.77 (0.73, 0.81)* 0.81 (0.77, 0.84)* 0.79 (0.76, 0.83)* 0.81 (0.78, 0.85)*


Table 5. Proportion of directional errors (%) in the task with and without the concurrent discrimination task in trials with and in trials without a distractor in eachgroup

Without distractor With distractor


Withdiscrimination


Withdiscrimination

Gap Overlap Gap Overlap Gap Overlap Gap Overlap

Control 0 0 1 1 12 5 15 7PD 0 0 1 1 11 7 14 12*


Table 6. Proportion of correct judgements (%; 95% confidence interval) in trials with and trials without a distractor in each group


Withoutdistractor

Withdistractor

Withoutdistractor

Withdistractor

Control 82 (74, 89) 76 (69, 84) 83 (77, 89) 75 (66, 84)PD 72 (67, 78)* 64 (57, 71)* 69 (64, 74)** 67 (59, 76)

*Difference between the groups, P < 0.05. **Difference between the groups, P < 0.01.



discrimination task (t78 = 1.91, P = 0.06). This effect was mainlyattributable to a significant Group · Task interaction. The discrimi-nation task increased the gain of saccades 0.03 (0.002, 0.05) more inthe PD group than in the control group (t77 = 2.00, P < 0.05). Noother interactions were significant.

Short-latency and long-latency saccades

Without the perceptual task, subjects in the PD group made, onaverage, 30% of all saccades at latencies below 176 ms. Thisproportion was 36% when saccades were made with the discrimina-tion task. The control group made only 20% of all saccades atlatencies below 176 ms, and the perceptual task did not change thisproportion. The proportions of short-latency saccades differedbetween the groups in the trials with the discrimination task(t35 = )2.62, P = 0.01). Figure 5 shows the association of saccadelatency and gain for each group in two-dimensional density plots.

Discussion

The effect of a perceptual discrimination task on the production ofreflexive saccades was assessed in a group of people with PD and acontrol group. The discrimination task reduced saccade latencies morein the PD group than in the control group, resulting in abnormallyshort mean reflexive saccade latencies in the PD group (Figs 2 and 5).The discrimination task increased saccade gain in both groups, butsaccades in the PD group were still abnormally hypometric ascompared with the control group (Fig. 3). Also, the performance of the

perceptual discrimination task was impaired in the PD group ascompared with the control group.

The latency of reflexive saccades in PD

The initiation of reflexive saccades at abnormally short latencies (orhyper-reflexivity) in the PD group is consistent with previous reportsof abnormal facilitation of the reflexive saccade system in PD (Briandet al., 2001; Armstrong et al., 2002; Kingstone et al., 2002; Chanet al., 2005; van Stockum et al., 2008). In our paradigm, the reductionof saccade latencies in the PD group resulted from top-downfacilitation of the saccade system, in response to the attentionaldemands of the discrimination task. The mean latency of reflexivesaccades was abnormally reduced in the PD group, especially in trialswith an overlap. This observation is consistent with the result of arecent comprehensive meta-analysis of investigations of reflexivesaccade latency in PD (Chambers & Prescott, 2010). This meta-analysis concluded that overall latencies of reflexive saccades areprolonged in PD, except in overlap paradigms and for targets ateccentricities of 5� and smaller (Chambers & Prescott, 2010).Chambers and Prescott’s hypothesis regarding the underlying neuralcause of this phenomenon involves altered retinal inputs into thesuperior colliculus resulting from dopamine depletion in PD. Thisinterpretation, however, does not explain the abnormal reduction ofsaccade latencies caused by task demands in the PD group in ourstudy. The association of short saccade latencies with smaller gainvalues of reflexive saccades was the same in both groups.

The gain of reflexive saccades in PD

Despite the reduction in latencies, members of the PD group were ableto increase the mean gain of their reflexive saccades in thediscrimination task. However, overall reflexive saccades remainedhypometric in our PD group as compared with the control group.Many investigations have found that, in contrast to the gain ofvoluntary saccades, gain values of reflexive saccades are normal in PD(Crawford et al., 1989; Lueck et al., 1992; Vidailhet et al., 1994;Shaunak et al., 1999). However, our results are consistent with some

Fig. 4. The scatterplot on the left shows the mean saccade latency on the x-axis and the proportion of correct discriminations for each subject on the y-axis. Thescatterplot on the right shows the mean saccade gain on the x-axis and the proportion of correct discriminations for each subject on the y-axis. In the control group,better performance in the discrimination task was associated with longer mean latencies. In the PD group, better performance in the discrimination task wasassociated with larger mean saccade gain values.

Table 7. Proportions of trials with saccade initiation before, during or afterdiscrimination symbol onset and the proportion of correct perceptual discrim-inations in each category

Saccade beforesymbol onset

Saccade duringsymbol onset

Saccade aftersymbol onset

Control 2% (86% correct) 12% (79% correct) 86% (78% correct)PD 5% (83% correct) 16% (67% correct) 79% (69% correct)



previous reports of hypometric reflexive saccades in PD (White et al.,1983; Rascol et al., 1989; Nakamura et al., 1991; Rottach et al.,1996; Armstrong et al., 2002). Saccadic hypometria in PD has beenattributed to striatal dopamine depletion and excessive inhibition in thesaccade system, but there is no strong evidence that levodopa therapyimproves saccade gain in PD (Nachev & Kennard, 2005).

Top-down effects

In tasks where visual input competes with top-down attentionalselection, people with PD often have difficulty in ignoring irrelevantvisual stimuli (Deijen et al., 2006; Machado et al., 2009). In ourparadigm, the distractor onsets, which occurred on a proportion oftrials, may have induced top-down inhibition, as subjects tried to avoiddistraction by the irrelevant onsets. However, the distractors affectedthe performance of the two tasks equally in both groups, and there wasno evidence that subjects in the PD group were more susceptible to theappearance of the distractors than subjects in the control group. Top-down inhibition, induced by the distractor onsets, would havecompeted with the task’s instructions, because the correct performanceof the task depended on subjects responding as quickly as possible tothe target onset. Instead, the instructions in this paradigm may haveinduced subjects to prepare for the onset of the target and the

discrimination symbol by increasing the size of the ‘attentionalwindow’ from the fixation area to encompass the potential targetlocations. The top-down strategy, to disengage attention from thefixation area before target onset, revealed abnormal facilitation of thesaccade system in the PD group. The finding that this facilitation wasmost obvious in the overlap condition suggests that it may beassociated with a reduction of fixation-related inhibition in the saccadesystem. In turn, reduced fixation-related inhibition may render theoculomotor system hypersensitive to visual inputs in PD. Thisinterpretation is also consistent with reports of increased bottom-updistractibility without impaired top-down distractor inhibition in PDand aged subjects (Langley et al., 1998; Troche et al., 2006). It couldbe proposed, consistent with the traditional assumption that saccadedeficits in PD reflect a failure of top-down control mechanisms, thatsubjects in the PD group were not able to use fixation-relatedinhibition to control the initiation of saccades in the discriminationtask. However, subjects in the control group did not increase theirlatencies in the discrimination task, but they reduced their saccadelatencies in overlap trials, and this effect was enhanced in the PDgroup.

Association of saccade latency, gain, and perceptualdiscrimination

In the PD group, the perceptual discrimination task facilitated theinitiation of saccades more than in the control group, but theperformance of the discrimination task was worse than in the controlgroup. In both groups, the gain of saccades initiated at latencies shorterthan 176 ms depended strongly on latency: shorter saccade latency wasassociated with smaller saccade gain. In the control group, betterperformance in the discrimination task was associated with longermean latency, but in the PD group, better performance in thediscrimination task was associated with larger mean gain values.Together, these results are consistent with previous studies showing thatlonger saccade latencies allow better pre-saccadic visual processing ofthe target stimulus and ensure better spatial accuracy of the eyemovement (Findlay, 1982; Findlay & Walker, 1999). If saccades aretriggered at very short latencies, not only the processing of visualinformation at the target location but also the gain of the saccade may bereduced (Ottes et al., 1985; Coeffe & O’Regan, 1987). This suggeststhat, in PD, hypometria of saccades may be associated with impairedprocessing of visual information at the saccade target location. In thePD group, the build-up of neural activity in the oculomotor systemduring saccade latency may have been insufficient to produce spatiallyaccurate saccades or to allow efficient perceptual discrimination.

Neurophysiology of the saccadic system in PD

From neurophysiology studies in monkeys, we know that saccades aretriggered when saccade neurons in the superior colliculus and thefrontal eye fields reach a threshold level of activity (Everling et al.,1998). The basal ganglia are preferentially involved in the initiation ofvoluntary saccades and in the prevention of unwanted saccades(Hikosaka et al., 2000). Basal ganglia output contributes to the controlof the saccade system by exerting a constant tonic inhibition via theSNr. When a saccade is to be made, a striatal signal selectivelyreleases this neural inhibition in the saccade system via the directpathway that connects the striatum directly with the SNr. Striataldopamine depletion interferes with this function, and results inexcessive inhibition (impaired release of inhibition) in the saccadesystem. During fixation, when saccades must be suppressed, theinhibitory output from the basal ganglia is enhanced via the indirect

Fig. 5. These two-dimensional density plots illustrate the effect of thediscrimination task on saccade latencies and gain in the PD and controlgroups. Each dot represents one saccade, with the latency shown on the x-axisand the gain on the y-axis. The plots in the column on the left represent saccadetrials without the discrimination task. The plots in the column on the rightrepresent saccade trials with the discrimination task. The top two rows showtrials with an overlap, and the bottom two rows show trials with a gap. The redlines show the association between gain and latency. The blue contour linesshow areas of equal frequency. Where the lines are close together, thefrequency changes rapidly, and where the lines are further apart, the frequencychanges more slowly. For saccades with latencies below 176 ms, gain increasedwith increasing latencies, but at latencies longer than 176 ms, saccades did notfurther increase in gain. The discrimination task promoted the production ofsaccades at latencies below 176 ms in the PD group, especially in the overlaptrials. In contrast, in the overlap trials, the control group produced a bimodallatency distribution, with a larger proportion of responses at latencies over176 ms than the PD group.



pathway that connects the striatum with the SNr via the externalcapsule of the globus pallidus and the subthalamic nucleus, and neuralactivity in the saccade system is further suppressed (Hikosaka et al.,2000). Together, these mechanisms allow the saccade system to selectappropriate saccades before any unwanted saccades are triggered. InPD, abnormally slow generation of voluntary saccades and abnormallyfast triggering of reflexive saccades may reflect pathological changesin the direct and the indirect pathways, respectively. PD and ⁄ or itstreatment may affect the ability to suppress neural activity in thesaccade system during fixation, which is normally associated withenhanced output from the SNr via the indirect pathway. Thisinterpretation is consistent with reports of impaired control of fixationin PD (Fielding et al., 2006; Pinnock et al., 2010).

Conclusion

We attribute the observed reduction of saccade latencies and the increasein saccade gain in the PD group to a top-down effect in response to thedemands of the discrimination task. This effect revealed a source ofabnormal facilitation of the saccadic system in the PD group. In the PDgroup, saccade latencies were abnormally short, and proportions ofexpress saccades and direction errors were increased, when saccadeswere made in conjunction with the discrimination task, especially inoverlap trials. The triggering of saccades at abnormally short latencies,especially in overlap trials, may reflect impairment of fixation-relatedinhibition in the saccade system in PD, which would result in anenhanced oculomotor response to visual inputs. In our paradigm, thisenhanced response to visual inputs promoted the initiation of saccades atshort latencies, but it did not benefit the performance of the discrim-ination task in the PD group. Diminished fixation-related inhibition maybe a direct result of pathology or a compensatory mechanism in PD.A potentially enhanced response to visual inputs is consistent withconverging evidence that, in PD, the allocation of visual attention maybe abnormally dominated by salient visual inputs. This phenomenon hasbeen observed in various paradigms, for instance in a set shifting task(Cools et al., 2010), a pro-anti-saccade switching task (Cameron et al.,2010), a manual response task (Deijen et al., 2006), a cueing task (Seiss& Praamstra, 2006), and a saccade taskwith distractors (Machado et al.,2009). This type of facilitation may adversely affect the performance ofsaccade tasks where visual input competes with top-down saccadeselection, such as an anti-saccade task, whereas it may enhance theperformance of saccade tasks such as the paradigm used in this study,where visual input reinforces top-down saccade selection.

Acknowledgements

The authors would like to thank the reviewers for commenting on earlierversions of the manuscript. S.v.S. was supported by a New Zealand TECScholarship.

Abbreviations

H&Y, Hoehn and Yahr; MDS-UPDRS, Movement Disorder Society-sponsoredrevision of the Unified Parkinson’s Disease Rating Scale; PD, Parkinson’sdisease; SNr, substantia nigra pars reticulata; SOA, stimulus onset asynchrony.

References

Amador, S.C., Hood, A.J., Schiess, M.C., Izor, R. & Sereno, A.B. (2006)Dissociating cognitive deficits involved in voluntary eye movement dysfunc-tions in Parkinson’s disease patients. Neuropsychologia, 44, 1475–1482.

Armstrong, I.T., Chan, F., Riopelle, R.J. & Munoz, D.P. (2002) Control ofsaccades in Parkinson’s disease. Brain Cognit., 49, 198–201.

Briand, K.A., Strallow, D., Hening, W., Poizner, H. & Sereno, A.B. (1999)Control of voluntary and reflexive saccades in Parkinson’s disease. Exp.Brain Res., 129, 38–48.

Briand, K.A., Hening, W., Poizner, H. & Sereno, A.B. (2001) Automaticorienting of visuospatial attention in Parkinson’s disease. Neuropsychologia,39, 1240–1249.

Cameron, I.G., Watanabe, M., Pari, G. & Munoz, D.P. (2010) Executiveimpairment in Parkinson’s disease: response automaticity and task switching.Neuropsychologia, 48, 1948–1957.

Chambers, J.M. & Prescott, T.J. (2010) Response times for visually guidedsaccades in persons with Parkinson’s disease: a meta-analytic review.Neuropsychologia, 48, 887–899.

Chan, F., Armstrong, I.T., Pari, G., Riopelle, R.J. & Munoz, D.P. (2005)Deficits in saccadic eye-movement control in Parkinson’s disease. Neuro-psychologia, 43, 784–796.

Chen, Y.F., Chen, T. & Tsai, T.T. (1999) Analysis of volition latency onantisaccadic eye movements. Med. Eng. Phys., 21, 555–562.

Chiu, G. (2010) CSIRO Mathematics, Informatics and Statistics bentcableAR:Bent-Cable Regression for Independent Data or Autoregressive Time Series,R package version 0.2.3. Available online at http://cran.r-project.org/web/packages/bentcableAR.

Coeffe, C. & O’Regan, J.K. (1987) Reducing the influence of non-target stimulion saccade accuracy: predictability and latency effects. Vis. Res., 27, 227–240.

Cools, R., Rogers, R., Barker, R.A. & Robbins, T.W. (2010) Top-downattentional control in Parkinson’s disease: salient considerations. J. Cogn.Neurosci., 22, 848–859.

Crawford, T., Henderson, L. & Kennard, C. (1989) Abnormalities ofnonvisually-guided eye movements in Parkinson’s disease. Brain, 112(Pt6), 1573–1586.

Deijen, J.B., Stoffers, D., Berendse, H.W., Wolters, E. & Theeuwes, J. (2006)Abnormal susceptibility to distracters hinders perception in early stageParkinson’s disease: a controlled study. BMC Neurol., 6, 43.

Deubel, H. (2008) The time course of presaccadic attention shifts. Psychol.Res., 72, 630–640.

Everling, S., Pare, M., Dorris, M.C. & Munoz, D.P. (1998) Comparison of thedischarge characteristics of brain stem omnipause neurons and superiorcolliculus fixation neurons in monkey: implications for control of fixationand saccade behavior. J. Neurophysiol., 79, 511–528.

Fielding, J., Georgiou-Karistianis, N. & White, O. (2006) The role of the basalganglia in the control of automatic visuospatial attention. J. Int. Neuropsy-chol. Soc., 12, 657–667.

Findlay, J.M. (1982) Global visual processing for saccadic eye movements. Vis.Res., 22, 1033–1045.

Findlay, J.M. & Walker, R. (1999) A model of saccade generation based onparallel processing and competitive inhibition. Behav. Brain Sci., 22, 661–674; discussion 674–721.

Goetz, C.G., Tilley, B.C., Shaftman, S.R., Stebbins, G.T., Fahn, S., Martinez-Martin, P., Poewe, W., Sampaio, C., Stern, M. B., Dodel, R., Dubois, B.,Holloway, R., Jankovic, J., Kulisevsky, J., Lang, A. E., Lees, A., Leurgans,S., LeWitt, P.A., Nyenhuis, D., Olanow, C.W., Rascol, O., Schrag, A., Teresi,J.A., van Hilten, J.J. & LaPelle, N. (2008) Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): scale presentation and clinimetric testing results. Mov. Disord., 23,2129–2170.

Gurvich, C., Georgiou-Karistianis, N., Fitzgerald, P.B., Millist, L. & White,O.B. (2007) Inhibitory control and spatial working memory in Parkinson’sdisease. Mov. Disord., 22, 1444–1450.

Hikosaka, O., Takikawa, Y. & Kawagoe, R. (2000) Role of the basal ganglia inthe control of purposive saccadic eye movements. Physiol. Rev., 80, 953–978.

Hood, A.J., Amador, S.C., Cain, A.E., Briand, K.A., Al-Refai, A.H., Schiess,M.C. & Sereno, A.B. (2007) Levodopa slows prosaccades and improvesantisaccades: an eye movement study in Parkinson’s disease. J. Neurol.Neurosurg. Psychiatry, 78, 565–570.

Kingstone, A., Klein, R., Morein-Zamir, S., Hunt, A., Fisk, J. & Maxner, C.(2002) Orienting attention in aging and Parkinson’s disease: distinguishingmodes of control. J. Clin. Exp. Neuropsychol., 24, 951–967.

Langley, L.K., Overmier, J.B., Knopman, D.S. & Prod’Homme, M.M. (1998)Inhibition and habituation: preserved mechanisms of attentional selection inaging and Alzheimer’s disease. Neuropsychology, 12, 353–366.

Le Heron, C.J., MacAskill, M.R. & Anderson, T.J. (2005) Memory-guidedsaccades in Parkinson’s disease: long delays can improve performance. Exp.Brain Res., 161, 293–298.

Lueck, C.J., Crawford, T.J., Henderson, L., Van Gisbergen, J.A., Duysens, J. &Kennard, C. (1992) Saccadic eye movements in Parkinson’s disease: II.



Remembered saccades – towards a unified hypothesis? Q. J. Exp. Psychol. AHum. Exp. Psychol., 45, 211–233.

Machado, L., Devine, A. & Wyatt, N. (2009) Distractibility with advancing ageand Parkinson’s disease. Neuropsychologia, 47, 1756–1764.

Mink, J.W. (1996) The basal ganglia: focused selection and inhibition ofcompeting motor progammes. Prog. Neurobiol., 50, 381–425.

Montagnini, A. & Chelazzi, L. (2005) The urgency to look: prompt saccades tothe benefit of perception. Vis. Res., 45, 3391–3401.

Nachev, P. & Kennard, C. (2005). Oculomotor dysfunction. In Pfeiffer, R.F. &Bodis-Wollner, I. (Eds.), Parkinson’s Disease and Nonmotor Dysfunction.Humana Press, Totowa, NJ, pp. 271–280.

Nakamura, T., Kanayama, R., Sano, R., Ohki, M., Kimura, Y., Aoyagi, M. &Koike, Y. (1991) Quantitative analysis of ocular movements in Parkinson’sdisease. Acta Otolaryngol. Suppl., 481, 559–562.

Nasreddine, Z.S., Phillips, N.A., Bedirian, V., Charbonneau, S., Whitehead, V.,Collin, I., Cummings, J.L. & Chertkow, H. (2005) The Montreal CognitiveAssessment, MoCA: a brief screening tool for mild cognitive impairment.J. Am. Geriatr. Soc., 53, 695–699.

Ottes, F.P., Van Gisbergen, J.A. & Eggermont, J.J. (1985) Latency dependenceof colour-based target vs nontarget discrimination by the saccadic system.Vis. Res., 25, 849–862.

Peirce, J.W. (2007) PsychoPy – psychophysics software in Python. J. Neurosci.Methods, 162, 8–13.

Peirce, J.W. (2008) Generating stimuli for neuroscience using PsychoPy. Front.Neuroinformatics, 2, 10.

Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & Team. (2009) nlme: Linearand Nonlinear Mixed Effects Models, R package version 3.1-96. Availableonline at: http://cran.r-project.org/web/packages/nlme.

Pinnock, R.A., McGivern, R.C., Forbes, R. & Gibson, J.M. (2010). Anexploration of ocular fixation in Parkinson’s disease, multiple system atrophyand progressive supranuclear palsy. J. Neurol., 257, 533–539.

Rascol, O., Clanet, M., Montastruc, J.L., Simonetta, M., Soulier-Esteve, M.J.,Doyon, B. & Rascol, A. (1989) Abnormal ocular movements in Parkinson’sdisease. Evidence for involvement of dopaminergic systems. Brain, 112(Pt5), 1193–1214.

Rottach, K.G., Riley, D.E., DiScenna, A.O., Zivotofsky, A.Z. & Leigh, R.J.(1996) Dynamic properties of horizontal and vertical eye movements inparkinsonian syndromes. Ann. Neurol., 39, 368–377.

Seiss, E. & Praamstra, P. (2006) Time-course of masked responsepriming and inhibition in Parkinson’s disease. Neuropsychologia, 44,869–875.

Shaunak, S., O’Sullivan, E., Blunt, S., Lawden, M., Crawford, T., Henderson,L. & Kennard, C. (1999) Remembered saccades with variable delay inParkinson’s disease. Mov. Disord., 14, 80–86.

van Stockum, S., MacAskill, M., Anderson, T. & Dalrymple-Alford, J. (2008)Don’t look now or look away: two sources of saccadic disinhibition inParkinson’s disease? Neuropsychologia, 46, 3108–3115.

Troche, S.J., Trenkwalder, C., Morelli-Canelo, M., Gibbons, H. & Rammsayer,T.H. (2006) Unimpaired negative but enhanced positive priming inParkinson’s disease: evidence from an identity and a location priming task.Neuropsychologia, 44, 1811–1821.

Trottier, L. & Pratt, J. (2005) Visual processing of targets can reduce saccadiclatencies. Vis. Res., 45, 1349–1354.

Uc, E.Y., Rizzo, M., Anderson, S.W., Qian, S., Rodnitzky, R.L. & Dawson,J.D. (2005) Visual dysfunction in Parkinson disease without dementia.Neurology, 65, 1907–1913.

Vidailhet, M., Rivaud, S., Gouider-Khouja, N., Pillon, B., Bonnet, A.M.,Gaymard, B., Agid, Y. & Pierrot-Deseilligny, C. (1994) Eye movements inparkinsonian syndromes. Ann. Neurol., 35, 420–426.

White, O.B., Saint-Cyr, J.A., Tomlinson, R.D. & Sharpe, J.A. (1983) Ocularmotor deficits in Parkinson’s disease. II. Control of the saccadic and smoothpursuit systems. Brain, 106, 571–587.



Documents

A perceptual discrimination task abnormally …nzbri.org/resources/publications/60/van_Stockum_European...A perceptual discrimination task abnormally facilitates reﬂexive saccades